[extropy-chat] Kurzweil article in new scientist

[Julian Assange]

>From a letter from Julian Assange to Chris Goddard:

If you like that kind of thing, I recommend a very funny book called
"Great Mambo Chicken and the Transhuman Condition" by Ed Regis. A lot of
these ideas have been blowing about the futurist set.

As for Ray, I don't trust him much, he's a bit of a hype merchant. It's
not clear to me that there is a law of accelerating returns, at least
not the one Kurzweil had traditionally suggested. Part of the problem is
that it is not clear what is the "natural scale" many these observations
(e.g cpu operations per dollar) should be on if you want to pull them
together and say something meaningful.

The only objective way I can think about accelerating returns is in
terms of energy and entropy measures. It seems clear that in the natural
world an organinism can gain some more energy from its environment by
spending some energy on computations in order to more efficaciously
navigate its destiny. But the gain is sub-linear. If we spend some
energy on N computations we rarely find an energy profit in proportion
to N. The relationship is not only strongly sub-linear, but is limited
to efficiency gains on energy inputs.

With the exception of revealing a path to new energy sources, which is
in some sense the whole deal, computation does not make energy, but
permits us to produce more entropy with the same energy. But even here
we can not say, forget energy --- that's just another unit, what we're
really interested is in entropy. It's entropy that measures the the
interesting things that life does. We can't say this because whilst one
can convert energy into entropy with greater efficiency as Moores law
goes on and buys us exponentially more computations per unit of energy
spent, we're not that far off the theoretical limit for the conversion
efficiency already. So the best way to increase entropy is to increase
energy inputs. i.e increase energy sources.  Four times the energy
inputs at current conversion efficiency will easily beat the maximum
possible conversion efficiency.

But notice that the total entropy production per second associated with
life on earth has been accelerating for billions of years. Restricting
our thoughts to man in the last 100 years, the exponential growth in
the economy is interdependent with an exponential growth in the
consumption of energy inputs, so most of the entropy acceleration is
from man pulling in more energy inputs from stored sources like oil,
coal and uranium and encircling the already exploited renewable inputs
of other life forms by... eating them.

But what of the computational factor? When you do something more
intelligently, i.e more efficiently than you once did, but still consume
the same energy input (by increasing production) you produce more
entropy than before. Infact we may define efficiency this way.
We can see this clearly in generators that reuse
their waste heat and emit it at a lower temperature and coprophagic
animals that decrease the energy density of their feces.

But this is what all of life is individually and as a whole; life is a
super enzyme that makes a fast path from low entropy states to high
entropy states. A blade of grass takes in sunlight at visible
frequencies then re-radiates at infrared frequencies (higher entropy)
after being eaten by a cow. If we include the act-ivies of man under
this definition of life, then, life is getting, well, livelier. Looking
at the entropy acceleration here on earth we might project our minds
forward and say we are on a mission to randomize the entire universe. To
boldly enropise where no one has entropised before.

The only problem with this vision is that when we look out on the
heavens we see stars. Lots of stars. Vast blobs of high energy density
shedding relatively low entropy radiation into the night. Why isn't life
elsewhere in the universe harnessing it? Sucking in the visible and re-
radiating in the far infrared. It has had, after all, plenty of time to
cover great swathes of galaxies if it had been onwards and outwards for
a billion years. Does life arise rarely, or do civilisations collapse
vigorously? Entropy always increases but its second derivative holds no
such guarantees. Perhaps the dark ages on earth explains the extent of
light we see in the heavens or do we just see what we see and the seeds
of
life have already consumed the twinking of the great inter cluster
voids?

On Wed, 28 Sep 2005 09:04:25 +1000 (EST),
f.Goddard at ms.unimelb.edu.au said:
> Hi people,
>
> There is an article you all must read in the latest New Scientist (the
> one dated 24th September), written by Ray Kurzweil.  It is an
> interesting perspective on how things might pan out in the next thirty
> or forty years.
>
> Cheers, Chris